The repostory includes a complete design of C0-compiler, fully coded in C language and referring to the most classical compiler architecture. Besides the modern basic architecture, the author employed several optimized methods to obtain more efficient ultimate MIPS code.
More than the code version, the repostory also contains a final concluding article about my design and plenty of typical testing code which is coded in C0 language, contributed by me and my classmates.
As we all know, a classical compiler mainly includes these important parts: Lexical Analysis, Grammar Analysis, Semantic Analysis & Intermediate Code Generation, Code Generation, Optimization and Operations of Symbol Table. A basic process example is shown as follows:
Some mission settings have to be mentioned in this section, which mainly are outlines of the course.
- Difficulty Division We have 3 different difficulty division in our course, which have been given as follows. We choose the Level III.
-
C0-Grammar: A Special Teaching Grammar
C0-Grammar is a C-likely language grammar, and it mainly abstract the essence part, for instance, recurrence, loop, etc.
＜加法运算符＞ ::= +｜-
＜乘法运算符＞ ::= *｜/
＜关系运算符＞ ::= <｜<=｜>｜>=｜!=｜==
＜字母＞ ::= ＿｜a｜．．．｜z｜A｜．．．｜Z
＜数字＞ ::= 0｜＜非零数字＞
＜非零数字＞ ::= 1｜．．．｜9
＜专用符号＞ ::= （｜）｜｛｜｝｜,｜;｜＝
＜字符＞ ::= ‘<加法运算符＞｜＜乘法运算符＞｜＜字母＞｜＜数字＞’
＜字符串＞ ::= "｛十进制编码为32,33,35-126的ASCII字符｝"
＜程序＞ ::= ［＜常量说明部分＞］［＜变量说明部分＞］{＜有返回值函数定义部分＞}{＜无返回值函数定义部分＞}＜主函数＞
＜常量说明部分＞ ::= const＜常量定义＞;{ const＜常量定义＞;}
＜常量定义＞ ::= int＜标识符＞＝＜整数＞{,＜标识符＞＝＜整数＞}
|float＜标识符＞＝＜实数＞{,＜标识符＞＝＜实数＞}
| char＜标识符＞＝＜字符＞{,＜标识符＞＝＜字符＞}
＜整数＞ ::= ［＋｜－］＜非零数字＞｛＜数字＞｝｜０
＜实数＞ ::= ［＋｜－］<整数>[.<整数>]
＜标识符＞ ::= ＜字母＞｛＜字母＞｜＜数字＞｝
＜声明头部＞ ::= int＜标识符＞ |float ＜标识符＞|char＜标识符＞
＜变量说明部分＞ ::= <变量定义>;{<变量定义>;}
＜变量定义＞ ::= ＜类型标识符＞＜标识符＞{,＜标识符＞}
＜类型标识符＞ ::= int | float | char
＜有返回值函数定义部分＞ ::= ＜声明头部＞‘(’ ＜参数＞ ‘)’ ‘{’ ＜复合语句＞ ‘}’
＜无返回值函数定义部分＞ ::= void<标识符>‘(’ <参数> ‘)’ ‘{’ <复合语句> ‘}’
＜复合语句＞ ::= ［＜常量说明部分＞］［＜变量说明部分＞］＜语句列＞
＜参数＞ ::= ＜参数表＞
＜参数表＞ ::= ＜类型标识符＞＜标识符＞{,＜类型标识符＞＜标识符＞}| 空
＜主函数＞ ::= void main ‘(’ ＜参数＞ ‘)’ ‘{’ ＜复合语句＞ ‘}’
＜表达式＞ ::= ［＋｜－］＜项＞{＜加法运算符＞＜项＞}
＜项＞ ::= ＜因子＞{＜乘法运算符＞＜因子＞}
＜因子＞ ::= ＜标识符＞｜‘(’＜表达式＞‘)’｜＜整数＞｜＜有返回值函数调用语句＞|<实数>|＜字符＞
＜语句＞ ::= ＜条件语句＞｜＜循环语句＞｜‘{’<语句列>‘}’｜＜返回语句＞;|＜有返回值函数调用语句＞; |＜无返回值函数调用语句＞；
｜＜赋值语句＞;｜＜读语句＞;｜＜写语句＞;｜＜空＞
＜赋值语句＞ ::= ＜标识符＞＝＜表达式＞
＜条件语句＞ ::= if ‘(’＜条件＞‘)’＜语句＞［else＜语句＞］
＜条件＞ ::= ＜表达式＞＜关系运算符＞＜表达式＞｜＜表达式＞ //表达式为0条件为假，否则为真
＜循环语句＞ ::= while ‘(’＜条件＞‘)’＜语句＞| for‘(’ ＜标识符＞＝＜表达式＞；<条件>；＜标识符＞＝＜表达式＞‘)’<语句>
＜有返回值函数调用语句＞ ::= ＜标识符＞‘(’＜值参数表＞‘)’
＜无返回值函数调用语句＞ ::= ＜标识符＞‘(’＜值参数表＞‘)’
＜值参数表＞ ::= ＜表达式＞{,＜表达式＞}｜＜空＞
＜语句列＞ ::= ＜语句＞｛＜语句＞｝
＜读语句＞ ::= scanf ‘(’＜标识符＞‘)’
＜写语句＞ ::= printf ‘(’[<字符串>,][＜表达式 ＞] ‘)’
＜返回语句＞ ::= return[‘(’＜表达式＞‘)’]
If need, you could constract this with the classical process mentioned before.
The mainly optimization methods the coder employ include Constant Combination, DAG graph calculation, counting to decide global register, peep optimization. In fact compiling optimization is really a challenging mission, even harder than the whole architecture. And optimization never ends.
Please click here to download the concluding document of the whole compiler design which is in .doc format. It includes the really detail designing thoughts and consideration, and includes some evaluation in the whole architecture. And of course the designer's journey of mind in the whole development.